Identifying and mitigating risk is a significant issue for companies today. One important aspect to identifying and mitigating risk is understanding how employees are using employer-provided computing devices. Accordingly, companies often monitor activity on these employer-provided devices. For instance, employers may monitor and evaluate search queries such as those input into an Internet search engine. In another example, employers may monitor email content of employees. Conventional systems of monitoring these activities include analyzing the content to determine whether any of the terms used match keywords identified as indicating risk. However, this system of identifying keywords results in many false positives because a word matching a keyword may be used as different parts of speech that may indicate different levels of risk. These conventional systems do not account for use of terms as different parts of speech and, thus, can provide inaccurate results.